Landsat-8 Operational Land Imager Change Detection Analysis Using Remote Sensing and GIS Techniques in Al-Manathera District, Najaf-Iraq

In this study, Al-Manathera district was monitored and mapped using Landsat-8 OLI imagery from the years 2013 to 2019 as well as evaluated land use and land cover changes in the district. Maximum like-hood technology was achieved to create the signature class of significant LULC category, barren land, agriculture land, orchards, water bodies and built-up area. After certifying suitable accuracy value for each classified image, a detail classification change detection analysis was performed. Image differencing statistical change detection technology, change dynamics analysis was also applied to assess the statistics of two images. According to remote sensing (RS) and geographic information system (GIS) techniques, the study was a try for monitoring the change in LULC patterns of Al-Manathera for the period 2013 to 2019. This study showed that agricultural land, built-up area, barren land and water bodies decreased by -7.91, -0.20 -1.76 and -0.08%, respectively, while orchards increased by 9.95%.


INTRODUCTION
Change detection is of challenging problem.Remote sensing is used for planning at local and regional level.Remote sensing provides the efficient and cheaper means of spatial and temporal classification of inland water studies (Gardelle et al., 2010;Prigent et al., 2012;Zhang et al., 2015).Landsat-8 Operational Land Imager; OLI superior design is useful for mapping application as compared to previous Landsat series (Irons J. R., Dwyer J. L. Barsi, J. A., 2012;Markham, B. L. et al., 2010;Pehlevan, N., Schott, J. R., 2011;U.S Geological Survey, 2012).Landsat-8 OLI is appropriate for land use/cover mapping (Czapla-Myers, et al., 2015;Flood, N., 2014;Jiag, P., Li, Feng, Z., 2014;Knight, E., Kvaran, G., 2014;Ke, Y., et al., 2015;Pervez, W., 2016;Morfitt., R., 2015;Markham, B., 2014;Roy D., et al., 2014.This paper presents a change detection study of Landsat-8 Operational Land Imager (OLI) data of the study area for the four seasons and for six different cases.The post classification techniquehas been used in this paper due to its advantages.Different change detection methods were used in the literature depending upon its application (Almutairi, A., Warner, T. A., 2010;Hecheltjen, A., Thonfeld, F., Menz, G., 2014.The objective of the paper was: (i) to evaluate SVM classification on OLI data for the four seasons; (ii) to evaluate post classification change detection analysis of OLI SVM classified data for the six cases

STUDY AREA AND DATA SETS
This paper describes change detection analysis of SVM classified OLI data for the four seasons.OLI data parameters of the study area is shown in Table 1.

OLI SVM Classified Data and Change Detection Analysis
OLI SVM classified data was used for change detection analysis of six cases.

Experimental Setup
The ROI were selected by using high resolution imagery and maps.Following

Case 1: Change Detection from Winter to Spring
Change detection matrix (Table 2) from winter to spring shows decrease of spatial distribution of bare land 66.99%, dam water 43.85 %, built-up land 50.95 and increase of mixed trees 9.2%, shallow water 40.4 % and shrub 514.4%. Figure 2 shows a change of category from dam water to channel water, dam water to bushes, and dam water to bare land.Similarly, a change of category from bare land to bushes, bushes to mixed trees, builtup area to bushes and built-up area to bare land resulted increase of bushes in spring from winter.Change detection from winter to spring resulted reduction in dam water mapping and increases of bushes.Change detection matrix (Table 3) shows decrease of spatial distribution of bare land 16.01%, dam water 32.52%, built-up land 40.3 and increase of mixed trees 22.1%, channel water 181.5% and bushes 121.5%. Figure 3 shows a change of category from dam water to channel water, dam water to bare land.Similarly, a change of category from bare land to bushes, built-up land to bare land, bushes to mixed trees resulted increases of vegetation in summer compared to winter.A change of category from mixed trees to dam water resulted near the shoreline.Change of category from bare land to built-up is due to seasonal variation.Change detection from winter to summer resulted reduction in dam water mapping and increase of vegetation.Change detection matrix (Table 4) from winter to autumn shows decrease of spatial distributions of mixed trees 90.81%, bare land 21.50% and increase of channel water 2.1 %, dam water 172.7%, bushes 94.3% and built-up land 16.72%.Figure 4 shows category changes with increase in dam water mapping from bare land to deep water and from bare land to bushes.Similarly, small category changes from built-up land to bare land result due to seasonal variations.Change detection matrix (Table 5) from spring to summer shows decrease of spatial distributions of mixed trees 88.35%, channel water 32.97%, bushes 72.75% and increase of bare soil 201.05%, built-up land 157.58% and dam water 288.69%.
Figure 5 shows category changes from mixed trees to bushes, bare land to bushes.Similarly category change from bushes to built-up land and bare land to built-up land, bushes to bare land, channel water to built-up land resulted due to decrease of vegetation.Category change from dam water to bare land resulted due to seasonal variation.Change detection from spring to summer resulted reduction of vegetation and shallow water.Change detection matrix (Table 6) from spring to autumn shows decrease of spatial distribution of mixed trees 91.59%, Channel water 27.29%, bushes 68.37% and increase of bare land 137%, deep water 385.85% and built-up land 137.96%.Figure 6 shows category changes with increases in dam water mapping from bushes to dam water, bare land to dam water in areas near the shoreline.Similarly change of category from bushes to bare land, built-up land to bare land, mixed trees to bushes resulted decrease of vegetation.Change of category from bare land to built-up land, bushes to built-up land resulted decrease of vegetation.Small category changes from bare land to bushes resulted due to seasonal variation.Change detection matrix from Summer to Autumn shows (Table 7) decrease of spatial distributions of bare land 21.01%, mixed trees 27.78% and increase of dam water 24.99%, channel water 8.48% and bushes 16.04%.Figure 7 shows a category changes with an increase in dam water mapping from bare land to dam water, channel water to dam water in areas near the shoreline.Similarly, category changes from bare land to bushes, built-up land to bushes, mixed trees to bushes resulted increase of bushes.Change of category from bare land to built-up area, built-up area to bare land, and bushes to bare land resulted due to seasonal variations.

Figure 2 .
Figure 2. Change detection results of OLI SVM classified data from winter to spring 3.4 Change Detection from Winter to Summer (Case 2)

Figure 3 .
Figure 3. Change detection results of OLI SVM classified data from winter to summer 3.5 Change Detection from Winter to Autumn (case 3)

Figure 4 .
Figure 4. Change detection results of OLI SVM classified data from winter to autumn

Figure 6 .
Figure 6.Change detection results of OLI SVM classified data from spring to summer 3.7 Change Detection from Spring to Autumn (case 5)

Figure 7 .
Figure 7. Change detection results of OLI SVM classified data from spring to autumn

Figure 7 .
Figure 7. Change detection results of OLI SVM classified data from from summer to autumn results of this study confirmed the potential utility of OLI data change detection analysis The OLI SVM classified data was successfully classified with regard to all six test classes (i.e., bare land, built-up land, mixed trees, bushes, dam water and channel water) after pre-processing and atmospheric correction.OLI SVM classified data resulted higher overall accuracy (more than 92%) and kappa coefficient and thus suitable for change detection analysis.The OLI SVM-classified data for the four seasons were used for change detection analysis of six cases.Case1: change detection from winter to spring resulted reduction in dam water mapping and increases of bushes.Case2: change detection from winter to summer resulted reduction in dam water mapping and increase of vegetation.Case3: change detection from winter to autumn resulted with increase in dam water mapping.Case 4 : Change detection from spring to summer resulted reduction of vegetation and shallow water.Case 5: change detection from spring to autumn resulted decrease of vegetation.Case 6: Change detection from summer to autumn resulted increase of bushes and vegetation.These results established that the new OLI technology, with its higher overall accuracy suitable for post classification change detection analysis.

Table 1 :
Imaging geometry conditions and scene center latitudes and longitudes for Landsat-8 OLI

Table 2 :
Change Detection Percentage Operational Land Imager Data from Winter to Spring Season

Table 3 :
Change Detection Percentage Operational Land Imager Data from Winter to Summer Season

Table 4 :
Change Detection Percentage Operational Land Imager Data from Winter to Autumn Season

Table 5 :
Change Detection Percentage Operational Land Imager Data from Spring to Summer Season

Table 6 :
Change Detection Percentage Operational Land Imager Data from Spring to Autumn Season

Table 7 :
Change Detection Percentage Operational Land Imager Data from Summer to Autumn Season